Background context: Surgical treatment of lumbar spinal stenosis is one of the most frequent procedures in spinal surgery. Nevertheless, the predictive quality of instruments like treadmill test and MRI has not yet been clearly defined for the lumbar spinal stenosis.

Purpose: Objective of this study was to verify correlations between treadmill test and MRI findings in the narrowest spinal segment.

Methods: In a prospective study (EBM-level 3) 25 patients with inpatient treatment for lumbar spinal stenosis were investigated between 01/2005 and 06/2005. Exclusion criteria were a reduced walking capacity for other reasons (e.g. gonarthrosis). Treadmill tests were performed following the protocol by Deen. Patients estimated their maximum walking distance before performing the tests. Area of the dural sac, area of the neuroforamina and disc height at the dorsal margin were examined by MRI for the narrowest spinal segment. VAS and ODI were used as clinical assessment instruments.

Outcome Measures: The median age of patients was 67 yrs (IQR 58–73 yrs) with 44 % females. The median distance reached in the treadmill test was 70 m (IQR 30–135 m), the median estimated maximum walking distance 200 m (IQR 100–300 m). In the narrowest spinal segment the median area of the dural sac was 91 mm2 (IQR 65–143 mm2), the median area of the neuroforamina 43 mm2 (IQR 36–51 mm2) and the median disc height 1,4 mm (IQR 0,9–2,5 mm). The median VAS was 7 (IQR 6,5–8) and the median ODI 33 (IQR 32–37).

Results: The distance reached in the treadmill test correlated with the maximum walking distance estimated by the patients (Spearman’s rho=0,62, p=0,001), area of the dural sac (rho=0,54, p=0,006) and disc height (rho=0,45, p=0,03), but not with area of the neuroforamina and VAS. VAS correlated with the disc height (rho=−0,6, p=0,002), but with no other MRI findings.

Conclusions: The distance reached in the treadmill test and MRI findings in the narrowest spinal segment predict the level of clinical symptoms. Possible reasons for the slightly limited predictive quality of MRI findings are that this study did not regard the time of development of the stenosis, that effects of multilevel stenosis were not considered, and that MRI findings in lying position without lordosis might differ from findings in standing position. However, the treadmill test is a valid and highly practicable pre-surgery diagnostic instrument in lumbar spinal stenosis.